How to teach an AI assistant to write tests and model security threats during coding

▤ Summary

This article shows how to train an AI assistant to write tests and support security threat modeling during coding. It walks through TDD workflows, OWASP-based threat modeling, and production deployment steps using proagents workflows in Cursor. It also demonstrates how to harden a login flow by replacing an insecure SQL query with parameterized access, bcrypt password checks, and JWT handling. The piece positions these workflows as practical tools for building more secure software with AI assistance.

▦ Linked Companies